I aborted 'rpm -ihv strace-4.5.20-r2.armv7a.rpm' after 2 hours on a beaglebone. When using the 'mmc' led-trigger the LED will light up solid for the full 2 hours. RP said "we patch the sync calls out the -native version" and I'm going to suggest doing the same for the target version.
The sync operations typically happen for database writes as well as when the files themselves are written to the disk. Such a long time is atypical in my experience. It would be interesting to know what it was doing that triggered this behavior -- or if a bug was hit. Note, the version of RPM we're using is getting quite old, so we will want to uprev it in the future as well.
I spoke with Koen about this and offered a few suggestions. The no fsync on file writes patch is already applied to the target version of RPM, so that should not be a factor. One item mentioned in our IRC conversation was that it's possible that the MMC suffered a bad block during the write process, causing the issue that was observed. I'll follow up and see if any more information is available.
I have been unable to reproduce the issue locally. I am dropping this to a minor bug. If someone else is able to reproduce this, please let me know. --Mark
Resolving as no further reports of issue have surfaced. Please reopen if there are additional symptoms.